Parametric and nonparametric tests Flashcards
What are parametric tests?
They are used when you can assume the data is normal.
What are non-parametric tests?
Make no assumptions about the distribution
What do we assume if we are using a parametric test?
- Normality
- Homogeneity of variance
- Linearity
- Independence
Which tests do we use to test normality?
Shapiro wilk
Smirnovv - large samples
What number indicates that the kurtosis is of normal distribution?
3 - Mesokurtic
What is the term for negative kurtosis?
Platykurtic
What is the term for positive kurtosis?
Leptokurtic
What can you assume if skewness is between -1 and +1?
Normality in data
When performing a Shapiro-Wilk to test normality, what P-value are you expecting for normality?
P > 0.05 - opposite to normal significance
If S-W test gives you a P > 0.05 outcome, what test would you use?
Parametric test as data is normal
If S-W test gives you a P < 0.05 outcome, what test would you use?
Non-parametric test as data is not normal
When do we use the Kolmogorov-Smirnov test?
If number in sample is greater than 2000
How do we test for equal variance?
Levenes test
How do we know if samples have equal variance?
P > 0.05 - opposite to normal significance
